Some notes on claims about Australia's crime rates

After the attack at the Aurora movie theater, there have been a lot of discussion about Australia's gun control laws.  For example, see this piece at the Huffington Post. 

A decade-long examination of the program in the journal "Injury Prevention" concluded that "chances of gun death in Australia dropped twice as steeply" after the program was implemented. A study by Harvard University in the Spring of 2011 suggested that the program helped reduce, either causally or directly, firearm deaths, gun-related suicides and accidental shootings. The Washington Post, summarizing many of the studies, concluded that there was "strong circumstantial evidence for the law's effectiveness." . . .

Of course, the Huffington Post ignores that more sophisticated research finds no benefit from these laws.  A paper by Lee and Suardi in Economic Inquiry in 2009 does an excellent job of looking at the issue.

Here is the actual data from Australia.  First note that gun ownership exhibits a very interesting pattern that isn't often acknowledged.  There was a large gun buyback in 1996 and 1997 that reduced gun ownership from 3.2 to 2.2 million guns.  But immediately after that gun ownership increased dramatically and is essentially back to where it was before the buyback.  Why is that important?  Well, if it is the number of guns that is important, you should initially see a large drop in suicides or crimes and then see it increasing.  Yet, in none of these data series do you observe that pattern.

For example, homicides didn't fall until eight years after the laws.  It is not clear what theory they have for why the long delay would occur.  Nor can I even find an acknowledgment of that long lag in the cited literature.   A more natural explanation for the drop at the eight year point would be the substantial increases in police forces that occurred at that time.

Crime can change for lots of reasons besides changes in gun control laws.  One way to get a handle on this is to see how Australian homicides are changing relative to other crime rates.  A single continuous crime data series isn't available, but the two diagrams below show how homicides is falling almost continuously as a fraction of violent crime.  If anything, the drop in homicides relative to overall violent crime was biggest in the 1970s.  

See also homicides relative to overall crime.  Again, it is very hard to see any benefit from the gun control laws.

UPDATE: Comments on the Huffington Post Show today.  Purely cross-sectional evidence in references to the UK, Germany, etc. doesn't mention that homicide rates were even lower relative to the US before they had gun control.  The graph that they showed across countries (at about 12:55) makes it look like the US has a similar murder rate to other countries such as Russia, Mexico, and other South American countries.  For further information see here and here.  

Finally, despite all the claims to the contrary, Australia didn't ban guns, though it did impose stricter criteria for people to own certain types of guns.

UPDATE 2: Ann Coulter has this interesting discussion of a New York Times piece by Rosenthal claiming "More Guns = More Killing":

Rosenthal also produces a demonstrably false statistic about Australia's gun laws, as if it's a fact that has been carefully vetted by the Newspaper of Record, throwing in the true source only at the tail-end of the paragraph:
"After a gruesome mass murder in 1996 provoked public outrage, Australia enacted stricter gun laws, including a 28-day waiting period before purchase and a ban on semiautomatic weapons. ... Since, rates of both homicide and suicide have dropped 50 percent ...," said Ms. Peters, who lobbied for the legislation." . . .
Whether or not the homicide rate went up or down in Australia as a result of strict gun control laws imposed in 1997 is a fact that could have been checked by Times researchers. But they didn't, because facts wouldn't have given them the answer they wanted.
Needless to say, the effect of Australia's gun ban has been extensively researched by Australian academics. As numerous studies have shown: After the gun ban, gun homicides in Australia did not decline any more than they were expected to without a gun ban.Thus, for example, according to the Australian Institute of Criminology, the homicide rate has been in steady decline from 1969 to the present, with only one marked uptick in 1998-99 -- right after the gun ban was enacted.
The showstopper for anti-gun activists like Ms. Rosenthal and Ms. Peters is the fact that suicides by firearm seemed to decrease more than expected after the 1997 gun ban.
But so did suicides by other means. Something other than the gun ban must have caused people to stop guzzling poison and jumping off bridges. (Some speculate that it's the availability of anti-depressants like Prozac.)
Curiously -- and not mentioned by Rosenthal -- the number of accidental firearms deaths skyrocketed after Australia's 1997 gun ban, although the law included stringent gun training requirements.
It turns out, until the coroner has certified a death as a "suicide," it's classified as "unintentional." So either mandatory gun training has led to more accidents, or a lot of suicides are ending up in the "accident" column.
Most pinheadedly, especially for a graduate of the Harvard Medical School, Rosenthal says: "Before (the gun ban), Australia had averaged one mass shooting a year. (Since then,) there have been no mass killings."
Mass murder is a rare enough crime that any statistician will tell you discerning trends is impossible. In this country, the FBI doesn't even track mass murder as a specific crime category. . . .
Totally unbeknownst to Elisabeth Rosenthal, Australian academics have already examined the mass murder rate by firearm by comparing Australia to a control country: New Zealand. (Do they teach "control groups" at Harvard?)
New Zealand is strikingly similar to Australia. Both are isolated island nations, demographically and socioeconomically similar. Their mass murder rate before Australia's gun ban was nearly identical: From 1980 to 1996, Australia's mass murder rate was 0.0042 incidents per 100,000 people and New Zealand's was 0.0050 incidents per 100,000 people.
The principal difference is that, post-1997, New Zealand remained armed to the teeth -- including with guns that were suddenly banned in Australia.
While it's true that Australia has had no more mass shootings since its gun ban, neither has New Zealand, despite continuing to be massively armed. . . .

The paper by McPhedran and Baker that Ann might be citing has some interesting facts.  I frequently hear about the mass public shootings in Australia prior to the 1996 law, but it turns out that there were only 4 of them between 1980 and 1996.  8 other attacks were "domestic" attacks.  If you look at total multiple victim-suicide attacks in Australia, the attacks are hardly eliminated after 1996.

Other:

A discussion of Europe is available here.

Mayor Bloomberg's proposals are discussed in the Wall Street Journal here.

So-called assault weapon bans are discussed here.

The importance of gun free zones is discussed here.

Other gun control laws that could impact these public mass shootings is discussed here.
See also ALER piece here.